摘要
物候和时相信息在农作物种植信息提取方面有十分重要的应用价值,利用多时相LandsatTM数据,结合冬小麦的波谱和时相信息,成功提取了北京地区的冬小麦种植信息。首先,选用了2003年4月7日、5月1日、5月25日、6月18日不同时相的4景TM卫星影像,计算了不同时相的NDVI时间谱图像数据;其次,结合北京地区农作物种植的实际情况,提取并分析了北京春夏季主要植被地物(冬小麦、苜蓿、苗圃、春玉米、树林等)的NDVI时间谱特征;第三,利用不同时相的NDVI图像数据,通过NDVI图像通道间的逻辑运算算法,成功提取了2003年北京地区的冬小麦种植信息,提取精度达到96.92%;最后,与2002年收割小麦的统计数据相对比,监测了北京各郊区县的冬小麦种植结构调整情况。结果表明,多时相、多光谱遥感数据在作物种植信息的监测中有十分明显的技术优势和重要的应用潜力。
Phenological and temporal information can improve crop classification precision greatly. In this paper, multi-temporal Landsat TM images were selected to extract winter wheat planting information in Beijing area. Firstly, four TM images were acquired on April 7, May 1, May 25, June 18, 2003, when winter wheat was at erecting, anthesis, grain filling and milking stage respectively. Secondly, considering the actual crops planting status in Beijing area, the multi-temporal NDVI curves of the main vegetations, including winter wheat, spring maize, clover, nursery, forest, were analyzed and extracted; Thirdly, a logical classification method was designed to extract winter wheat planted area, and the area and distribution of winter wheat harvested in 2003 was successfully remotely sensed with a precision of 96.92%. Finally, the remotely sensed results were compared with the statistical data of winter wheat reaped in 2002, and the winter wheat planting structural adjustment of each county in Beijing was analyzed. The result indicated that multi-temporal and multi-spectral remotely sensed data had obvious technological advantage and important application potentiality in monitoring crop-planting information.
出处
《遥感技术与应用》
CSCD
2005年第2期238-242,共5页
Remote Sensing Technology and Application
基金
国家863项目(2003AA209040)
北京市自然基金项目(6021002)共同资助。
